Voice interactions for AI glasses
Start from the existing glasses template and its supported toolkit. Identify whether the glasses have a display before designing the interaction. Keep voice responses brief; a spoken confirmation should name the action and outcome.
Provide clear listening, working, cancelled and failed states. Support an interrupt or cancel command. Never record in the background or save camera or microphone data without a deliberate user action. Keep the phone usable when glasses disconnect.
Test a denied microphone permission, missing network, unavailable speech recognition and a noisy environment. A browser emulator validates the flow; it does not prove real-device voice support. Experimental toolkit capabilities require the device's developer setup. Never claim that an emulated input was tested on glasses.
Read the Graspable guide for the current workflow.